Tomography: Its relevance and relative errors

In addition to basic physics and biology problems, and problems with the computation of high-LET radiation effects, detailed patient specification must be considered for treatment-planning purposes. This is more critical in high-LET work because of sharp dose gradients; for example, 60Co therapy central-axis dose fall-off is roughly 5% per centimeter, whereas fall-off could be as great as a 90% dose effect with pion therapy. Recent advances in tomography offer the promise of reaching the level of patient specification necessary for treatment planning. Conventional radiation therapy patient specification has been shown to have errors on the order of 2 cm, whereas computer-assisted tomography (CAT) has a resolution of about 2 mm. However, registration of CAT information for a patient in the treatment position may be the limiting factor (on the order of 0.5 cm). This technique will not account for patient movement, changes in patient definition due to bodily function, or changes due to irradiation over the course of treatment. Electrofluorotomography may solve some of these problems. This approach may provide registration information analogous to that currently provided by therapy simulators and verification analogous to therapy port films. Both of these tomography systems measure the dose absorbed through the patient from a low-energy X-ray spectrum. This has been shown to correlate best with electron density, but there are striking anomalies. It may be difficult to provide information on physical density or atomic number. Several radiation therapy groups are using CAT information directly for treatment planning. Improvements over conventional patient specification are highly site-specific. For the majority of treatments, the local efficiency and non-uniformity factors do not change in CAT-assisted plans.